If poorer families with children aged from 3 to 8 want larger and better dwelling units, then for a constant housing budget they will likely have to locate in worse neighborhoods with worse schools because rents are lower there. If the same families seek better neighborhoods or schools, then this will likely entail a smaller or otherwise worse unit because rents are higher in these better locations. This trade-off motivates the proposed study. It has three aims: (1) To understand the kinds of housing options available to families of different racial/ethnic and income groups, whose housing options are also likely to entail systematic variation in the quality of neighborhoods and schools;(2) To learn how these housing options and their links to neighborhoods and schools jointly affect the socioemotional development, academic achievement and health of children aged from 3 to 8 over a period of 40 months. This will allow us to conclude which combinations of housing, neighborhood and school quality best support positive child development. (3) We will also learn which specific combinations of housing, neighborhoods and schools families actually prefer to live in and thus the extent to which these preferences entail the best possible developmental outcomes for children of a given family income. The proposed study takes place in Cleveland, Dallas, Denver and Seattle, all chosen because they use a lottery to distribute their Section 8 housing vouchers and because they also differ on other factors that might affect how housing trade-offs impact on children - e.g., their racial/ethnic mix, the tightness of the private market for rental housing, ad how hard the foreclosure crisis hit. We seek to identify patterns of housing availability, child effects and housing preferences that are robust across cities. These provide the best foundations for future research and policy. A three-wave longitudinal survey will be conducted over 40 months. It will have two strata. The first is a randomly selected stratum of families in each city whose income is below the area median and who have at least one child aged from 3 to 8. The second stratum is of families with the same-aged children who have applied for a housing voucher and been declared eligible. Half will win a voucher and half not, with random deletion being used to prune study and control children to the numbers desired by our power analyses-- 3,450 children and 2,654 caregivers across strata and half of this in each of the separate strata. The child outcomes at three time points are: internalizing and externalizing behaviors, performance on math and reading tests and obesity. The analysis of Aim 1 depends on hedonic regressions that predict housing costs from attributes of dwelling units, neighborhoods and schools.
Aim 2 uses propensity scores with the general stratum and random assignment as an instrumental variable with the housing voucher stratum.
Aim 3 depends on a vignette experiment that varies three attributes for each of housing, neighborhood and school quality.

Public Health Relevance

The proposed research will improve the design of US housing policy for poorer families by providing policy makers with empirical evidence about: (1) how low income constrains available housing options, (2) how these housing options affect child development, and (3) how preferences and social-structural constraints jointly shape housing conditions among low income families.

National Institute of Health (NIH)
Eunice Kennedy Shriver National Institute of Child Health & Human Development (NICHD)
Research Project (R01)
Project #
Application #
Study Section
Special Emphasis Panel (ZRG1-SSPS-H (09))
Program Officer
Bures, Regina M
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Northwestern University at Chicago
Other Health Professions
Organized Research Units
United States
Zip Code